By Stephen Grossberg (auth.), Roberto Pirrone, Filippo Sorbello (eds.)

This ebook constitutes the refereed complaints of the twelfth overseas convention of the Italian organization for man made Intelligence, AI*IA 2011, held in Palermo, Italy, in September 2011. The 31 revised complete papers awarded including three invited talks and thirteen posters have been rigorously reviewed and chosen from fifty eight submissions. The papers are prepared in topical sections on computer studying; dispensed AI: robotics and MAS; theoretical matters: wisdom illustration and reasoning; making plans, cognitive modeling; typical language processing; and AI applications.

Show description

Read Online or Download AI*IA 2011: Artificial Intelligence Around Man and Beyond: XIIth International Conference of the Italian Association for Artificial Intelligence, Palermo, Italy, September 15-17, 2011. Proceedings PDF

Best international books

The Small-Mart Revolution: How Local Businesses Are Beating the Global Competition (BK Currents (Paperback))

Defenders of globalization, unfastened markets, and loose exchange insist there is not any substitute to mega-stores like Wal-Mart -- Michael Shuman begs to vary. In "The Small-Mart Revolution, Shuman makes a compelling case for his replacement enterprise version, one within which groups benefit from "going neighborhood" in 4 key spending different types: items, prone, power, and finance.

Additional info for AI*IA 2011: Artificial Intelligence Around Man and Beyond: XIIth International Conference of the Italian Association for Artificial Intelligence, Palermo, Italy, September 15-17, 2011. Proceedings

Example text

Advances in NIPS 12(3), 547–553 (2000) 14. : In defense of one-vs-all classification. Journal of Machine Learning Research 5, 101–141 (2004) 15. : Probabilistic discriminative kernel classifiers for multi-class problems. , Florczyk, S. ) DAGM 2001. LNCS, vol. 2191, pp. 246–253. Springer, Heidelberg (2001) 16. : Learning with Kernels. MIT Press, Cambridge (2001) 17. : Probability estimates for multi-class classification by pairwise coupling. Journal of Machine Learning Research 5, 975–1005 (2004) 18.

Semi-Supervised Multiclass Kernel Machines with Probabilistic Constraints 25 label fitting with a squared loss V (fj (xi ), yij ) = (fj (xi ) − yij )2 . Note that using a hinge loss leads to the same classification accuracies, as investigated in [14], and it would not make any substantial differences with respect to the selected V due to the nature of the constraints that we will introduce in the following (that will enforce fj in [0, 1]). In its unconstrained and fully Supervised formulation, the OVA scheme does not guarantee that the output values f1 (x), .

As for document representation, we adopted the bag of words approach, a typical method for representing texts in which each word from a vocabulary corresponds to a feature and a document to a feature vector. In this representation, all non-informative words such as prepositions, conjunctions, pronouns and very common verbs are disregarded by using a stop-word list. Moreover, the most common morphological and inflexional suffixes are removed by adopting a standard stemming algorithm. After determining the overall sets of features, their values are computed for each document resorting to the well-known TFIDF method.

Download PDF sample

Rated 4.61 of 5 – based on 39 votes